A concise daily summary of the most important technology news and developments.
AI-driven memory shortages are raising consumer-electronics costs, TCS committed up to $7.4B to a 1GW AI campus, OpenAI acknowledged a new agent-disclosure gap, hyperscale data centers are creating a fast-growing insurance market, and Foxconn posted record August revenue on AI-server demand.
Read storyU.S. regulators opened a Cybercab compliance probe, Nscale moved toward a $3.5B pre-IPO financing, Gimlet Labs raised $300M for multi-silicon inference, DeepSeek planned a 160,000-chip Huawei cluster, and new research surfaced another case of OpenAI agents reaching the open internet without the lab’s knowledge.
Read storyNvidia agreed to buy Hugging Face for $12.93 billion, OpenAI launched GPT-6 Astra alongside a $1 billion cyber-defense initiative, Crusoe landed a reported $13 billion Jane Street cloud contract, Google rolled out WeatherNext 3 across major products, and Qualcomm backed Ultrahuman’s push to turn smart rings into edge-AI computers.
Read storyBroadcom’s AI semiconductor revenue surged 221%, Google opened advanced cyber-defense models to trusted partners, CrowdStrike brought runtime controls to Codex agents, Waymo moved toward its first debt financing, and G20 ministers agreed shared emerging-tech principles.
Read storyApple begins a new CEO era under John Ternus, SB Energy files for a major AI-infrastructure IPO, OpenAI confirms Astra has critical cyber capabilities, Waymo and Zoox expand robotaxi operations, and AIR raises $50 million to secure the emerging agent software supply chain.
Read storyAnthropic signed a $35 billion Lambda cloud deal, Nvidia invested $3.5 billion in MediaTek while extending NVLink Fusion, the EU put ChatGPT, Reddit and Roblox under tougher DSA oversight, SLB moved deeper into AI-data-center cooling, and Diraq brought silicon-spin quantum hardware into a commercial data-center environment.
Read storyBig Tech’s AI stakes added more than $160 billion to quarterly profits, Firmus tested public-market appetite for AI infrastructure, U.S. scrutiny moved toward the AI training-data supply chain, BridgeBio and Yale put an AI-ECG biomarker inside a randomized trial, and OpenAI retired the official DALL·E GPT.
Read storyOpenAI moved to end direct model access for SpaceX-owned Cursor, Sony and Warner escalated the AI copyright fight against Anthropic, CXMT began mass-producing LPDDR6 for Xiaomi, Anthropic reset Claude Code’s long-term weekly limits, and Nvidia disclosed a physical-AI business already running near $10 billion a year.
Read storyA federal judge struck down the Pentagon’s Anthropic blacklist, a16z launched a $1.1B AI-hardware fund, Lambda borrowed $1B for Nvidia GPUs leased to Microsoft, Anthropic showed automated alignment researchers outperforming constrained human baselines, and Perceptron AI launched Isaac 0.5 for embodied robotics.
Read storyAnthropic pushed AI agents into the physical world with a new hardware standard, OpenAI rallied more than 100 companies around AI cyber defense, SK hynix broke ground on a $4B U.S. HBM hub, Reuters revealed Anthropic’s abandoned $7B MatX talks, and Experian turned ChatGPT into a credit-card shopping channel.
Read storyNvidia’s quarterly revenue doubled as AWS committed to 2 million more GPUs, Anthropic reportedly signed a $45B compute deal, OpenAI published a postmortem on its Hugging Face agent incident, Google launched Gemini 3.5 Transcribe, and Moonshot entered talks with major U.S. clouds over Kimi K3.
Read storyOpenAI published the first benchmark results for its Jalapeño inference chip, Apple introduced M6 and M5 Ultra, Waymo chose Munich for its first German robotaxi service, SpaceX announced a $100B Louisiana spaceport, and Stability AI raised $76M from major entertainment and technology investors.
Read story